Systems and Methods for Computing Mathematical Functions

1. An apparatus for computing mathematical functions, the apparatus comprising: a single hardware pipeline configured to perform a polynomial approximation of second degree or higher, the single hardware pipeline including a plurality of stages; and a plurality of data tables, each associated with at least one of an RCP (reciprocal), SQRT (square root), EXP (exponential) and LOG (logarithm) function and configured to be coupled to the single hardware pipeline according to at least one opcode, wherein each of the plurality of data tables includes data associated with implementing the associated function, and wherein the single hardware pipeline is further configured to compute at least one of the RCP, SQRT, EXP and LOG functions; wherein the polynomial approximation is of the form F(x)=a+b*(x−x0)+c*(x−x0)*(x−x1), where F(x) is an intermediate value used in the single hardware pipeline, x is an operand, x0 is a value greater than the operand, x1 is a value smaller than the operand, and values a, b, and c correspond to the values x0 and x1 and are retrieved from the plurality of data tables as coupled to the single hardware pipeline according to the at least one opcode.

2. The apparatus of claim 1 , wherein the polynomial approximation is a quadratic approximation.

3. The apparatus of claim 1 , wherein at least one of the plurality of stages is operable to be coupled to at least one of the plurality of data tables.

4. The apparatus of claim 1 , wherein each of the plurality of stages is operable to compute at least one term of the polynomial approximation.

5. The apparatus of claim 1 , wherein a first stage is operable to convert a floating point representation to a fixed point representation.

6. The apparatus of claim 1 , wherein a second stage is operable to convert a fixed point representation to a floating point representation.

7. A method comprising: identifying a first data table corresponding to at least one of a plurality of data tables, each data table associated with at least one of an RCP (reciprocal), SQRT (square root), EXP (exponential) and LOG (logarithm) function, wherein each of the data tables includes data associated with implementing the associated function; identifying a second data table corresponding to at least one of a plurality of data tables, each data table corresponding to at least one of the RCP, SQRT, EXP and LOG functions; performing, in a single hardware pipeline, a first polynomial approximation based on the first data table; performing, in the single hardware pipeline, a second polynomial approximation based on the second data table; and summing, in the single hardware pipeline, the first and second polynomial approximations to generate an output associated with at least one of the RCP, SQRT, EXP and LOG functions; wherein the first and second polynomial approximations are of the form F(x)=a+b*(x−x0)+c*(x−x0)*(x−x1), where F(x) is an intermediate value used in the single hardware pipeline, x is an operand, x0 is a value greater than the operand, x1 is a value smaller than the operand, and values a, b, and c correspond to the values x0 and x1 and are retrieved from the plurality of data tables as coupled to the single hardware pipeline according to the at least one opcode.

8. The method of claim 7 wherein the first polynomial approximation is of second degree or higher.

9. The method of claim 7 wherein the second polynomial approximation is of second degree or higher.

10. The method of claim 7 wherein the first polynomial approximation is a quadratic approximation.

11. The method of claim 7 wherein performing a first polynomial approximation includes a plurality of stages, each of the plurality of stages computing at least one term of the first polynomial approximation.

12. The method of claim 7 wherein performing a first polynomial approximation includes a plurality of stages, each of the plurality of stages computing at least one term of a quadratic approximation.

13. The method of claim 7 wherein performing a first polynomial approximation includes a plurality of stages, at least one of the plurality of stages converting a floating point representation to a fixed point representation.

14. The method of claim 7 wherein performing a first polynomial approximation includes a plurality of stages, at least one of the plurality of stages converting a fixed point representation to a floating point representation.